Using AI to Transform Operations

A practical guide for facilities and operations leaders

The goal of AI in operations isn't just 'smarter buildings'—it's simpler days for the people running them, coupled with better insights.

Executive Summary

Operation leaders have the opportunity to use artificial intelligence to reshape how their organizations manage facilities, assets, and decision-making. Yet many struggle to adopt it in ways that solve real problems rather than create new complexities.

At FMX we aim to use AI to provide a "comfortable interface" that reduces complexity, empowers your team, and transforms data from the field into actionable insights—without requiring your staff to become data scientists.

These insights help you transform your operations to better support your organization's mission.

Here's an expanded version with specific examples of operational transformation:

These insights help you transform your operations to better support your organization's mission through proactive resource allocation, smarter budget planning, enhanced stakeholder service, improved compliance and safety, and knowledge preservation.

AI Fundamentals

Before diving deeper into AI applications in operations, it's helpful to understand some fundamental concepts and terminology.

What is Artificial Intelligence (AI)?

At its core, artificial intelligence refers to computer systems that can perform tasks typically requiring human intelligence—such as understanding language, recognizing patterns, making decisions, and solving problems.

What is Machine Learning?

Machine learning is a subset of AI where systems learn from data rather than following explicitly programmed rules. Instead of coding "if this, then that" for every scenario, you provide examples and the system identifies patterns.

For example, rather than programming rules for identifying a maintenance issue, you'd show the system thousands of historical work orders, and it learns to recognize patterns that indicate specific types of problems.

What is Generative AI?

Generative AI is a recent breakthrough in machine learning that can create new content—text, images, code, or other outputs—based on patterns learned from training data.

Unlike traditional AI that classifies or predicts, generative AI can compose new work orders, draft reports, answer questions in natural language, or generate maintenance recommendations. This is the technology behind conversational assistants like FMX's Iris.

How is AI Different from a Standard Web Application?

Traditional web applications follow predetermined logic:

• You click a button → the application executes a specific function
• You enter data in a form → it gets stored in a database exactly as entered
• You run a report → it follows a predefined template

AI-enabled applications add adaptive capabilities:

• You describe what you need → the AI interprets your intent and takes appropriate action
• You enter data conversationally → the AI extracts relevant information and structures it appropriately
• You ask a question → the AI generates a custom response based on your specific context

The key difference: traditional applications require you to learn their language and processes; AI-enabled applications learn yours.

Key AI Capabilities Relevant to Operations

Natural Language Processing (NLP)

The ability to understand and generate human language. This enables conversational interfaces, automated categorization of work orders, and intelligent search.

Pattern Recognition

Identifying trends and anomalies in data that might not be obvious to humans. This powers predictive maintenance, resource optimization, and anomaly detection.

Automation

Executing routine tasks without human intervention. This includes routing work orders, scheduling maintenance, and generating reports.

Recommendation Systems

Suggesting optimal actions based on historical data and current context. This helps with prioritization, resource allocation, and decision support.

What AI Is Not

It's equally important to understand AI's limitations:

• AI is not sentient – It doesn't "understand" in the human sense; it identifies patterns in data
• AI is not infallible – It can make mistakes, especially with unusual situations or poor-quality data
• AI is not a replacement for expertise – It's a tool that augments human decision-making, not a substitute for professional judgment
• AI is not magic – It requires good data, clear objectives, and thoughtful implementation to deliver value

The Evolution: Why AI Finally Got Good Enough

From AI Winters to AI Spring

Artificial intelligence has experienced several "winters"—periods when grand promises failed to materialize and funding dried up. The term "artificial intelligence" itself was coined by John McCarthy in 1955, yet for decades, AI remained largely confined to research labs and narrow applications.

What changed? Three critical pillars finally converged:

1. The Big Data Explosion

Early AI was like a high-performance engine with no fuel. Machine learning models require massive amounts of data to identify patterns. Today, organizations generate quintillions of bytes of data daily through connected systems, IoT sensors, and digital workflows. This provides the "fuel" AI needs to learn.

2. Specialized Hardware (GPUs)

For decades, AI ran on standard processors that handled tasks sequentially. The breakthrough came when researchers discovered that Graphics Processing Units—originally designed for video games—are incredibly efficient at the parallel processing AI requires. Tasks that would have taken a year to calculate in the 1990s can now be completed in days.

3. The Transformer Breakthrough

Before 2017, AI struggled with context. The introduction of the Transformer architecture (the "T" in ChatGPT) allowed AI to process information holistically and assign attention to the most relevant elements. This is why modern AI can engage in human-like conversation, understand nuanced requests, and generate contextually appropriate responses.

What This Means for Operations

For facilities and operations leaders, these advances translate into practical capabilities that were impossible just a few years ago:

• Natural language interfaces that eliminate complex software navigation
• Predictive insights from maintenance data that previously sat unused
• Automated workflows that reduce administrative burden
• Real-time decision support that helps prioritize limited resources

The Problem: More Functionality, More Complexity

The Data Collection Paradox

Modern operations management systems offer unprecedented functionality. You can track work orders, manage assets, monitor energy consumption, schedule preventive maintenance, and generate detailed reports. Yet this power comes with a price: complexity.

The reality many operations leaders face:

• Data doesn't get collected because entering it is cumbersome and time-consuming
• Software feels unnatural requiring staff to navigate multiple screens and remember specific workflows
• Insights remain locked away because generating meaningful reports requires technical expertise
• Teams default to workarounds like spreadsheets and email because they're more "comfortable," even if less effective

This creates a vicious cycle: without accurate data, you can't generate insights. Without insights, you can't demonstrate value. Without demonstrated value, you can't justify investment in better tools or additional staff.

The Mobile Maintenance Challenge

Every unnecessary click, every mandatory field, every navigation step between screens creates friction that slows your team down. Multiply that friction across hundreds of work orders, dozens of staff members, and thousands of assets, and the impact compounds significantly.

This friction is especially problematic for maintenance technicians working in the field. Away from their desks and computers, technicians must rely on mobile devices to enter critical information. But mobile applications present their own unique challenges: navigating robust software on small screens, attempting to tap precisely while wearing gloves, struggling with touchscreens that don't respond to cold or sweaty hands, and trying to interact with devices after their hands are dirty from performing maintenance work.

The result? Frontline staff whose expertise should be focused on solving problems end up fighting both the complexity of the software and the physical constraints of their work environment.

The Solution: AI as the "Comfortable Interface"

Reframing the Role of AI

The promise of AI in operations isn't about replacing human expertise. It's about removing friction between people and systems. AI serves as a translation layer—a comfortable interface that allows staff to interact with complex systems using natural language and intuitive interactions.

"I don't want to work to make software work for me. I want my software to tell me it's got a problem. I'm trying to find ways to reduce my time in front of a computer. I'm much more effective in the field than I am behind a desk.”

Chris Bozarth Director of Maintenance/Facilities Owensboro Board of Education

This approach centers on three fundamental capabilities:

1. AI Helps Users Collect Accurate Data

The first barrier to better operations is getting good data into your systems. AI addresses this by:

Making data entry conversational

• Instead of navigating forms, staff can describe issues in plain language and use voice-to-text
• AI extracts relevant details and populates the right fields automatically
• Natural language reduces training time and increases adoption

Providing intelligent assistance

• Recommending relevant assets based on issue descriptions – When a user describes a problem, AI can suggest which specific assets might be affected based on similar past incidents.
• Auto-categorizing work orders based on historical patterns – AI analyzes past work orders to automatically assign appropriate categories, priorities, and tags to new requests, ensuring consistency and reducing manual classification effort.
• Flagging incomplete or inconsistent information before it causes problems – The system identifies missing critical details, conflicting data, or unusual patterns in real-time, prompting users to correct issues during data entry rather than discovering them later.

Meeting people where they are

• Accepting input through various channels (voice, text, mobile)
• Adapting to different terminology and communication styles
• Reducing the learning curve for new staff or infrequent users

2. AI Takes Action and Automates Using That Data

Once data exists in your system, AI can leverage it to:

Generate insights without manual analysis

• Identifying maintenance patterns that indicate emerging issues
• Highlighting resource allocation opportunities
• Surfacing anomalies that deserve attention

Automate routine decisions

• Routing work orders to appropriate staff based on skills and availability
• Prioritizing requests based on urgency, impact, and resource constraints
• Scheduling preventive maintenance at optimal intervals

Provide decision support for complex situations

• Answering questions about asset history, warranty status, or compliance requirements
• Recommending solutions based on similar past situations
• Generating reports and summaries in natural language

3. Actionable Insights That Transform Operations

AI-driven insights can help you create tangible operational improvements:

Proactive Resource Allocation

AI analyzes patterns in work orders, asset usage, and staff activity to help you deploy resources where they'll have the greatest impact:

• Identify which assets require the most attention and allocate preventive maintenance resources accordingly
• Recognize seasonal or cyclical patterns in service requests to optimize staffing levels
• Detect emerging issues before they become emergencies, allowing you to address problems during scheduled maintenance rather than costly after-hours responses
• Balance workloads across teams based on skills, availability, and historical performance data

Smarter Budget Planning

AI-generated insights provide the evidence you need to make compelling budget requests and strategic investment decisions:

• Forecast future maintenance costs based on asset age, usage patterns, and historical repair data
• Identify which assets are consuming disproportionate resources and may warrant replacement rather than continued repair
• Quantify the impact of deferred maintenance to justify capital investments
• Track actual versus budgeted spending in real-time with automated variance analysis
• Generate data-driven ROI projections for proposed equipment purchases or system upgrades

Enhanced Stakeholder Service

Better data and faster insights translate directly into improved service for the people who depend on your facilities:

• Reduce response times by automatically routing requests to the right person with the right skills
• Provide accurate status updates and completion estimates based on historical performance
• Anticipate needs before they're reported by identifying patterns that indicate developing issues
• Demonstrate accountability through transparent reporting on service levels and response times
• Close the communication loop by automatically notifying requesters when work is completed

Improved Compliance and Safety

AI helps ensure nothing falls through the cracks when it comes to regulatory requirements and safety protocols:

• Automatically schedule and track required inspections, certifications, and compliance activities
• Flag assets or systems approaching compliance deadlines before violations occur
• Identify safety patterns or recurring hazards that warrant systematic intervention
• Generate audit-ready documentation and compliance reports with minimal manual effort
• Ensure consistent application of safety protocols across all facilities and staff

Knowledge Preservation

Your organization's operational expertise shouldn't walk out the door when experienced staff retire or move on:

• Capture tribal knowledge by documenting solutions to recurring problems as they're resolved
• Build an institutional memory that makes asset history, vendor relationships, and past decisions accessible to current staff
• Accelerate onboarding for new team members by providing AI-powered access to historical context and best practices
• Identify your most effective troubleshooting approaches and make them available organization-wide
• Create continuity across leadership transitions by maintaining comprehensive operational records

These insights don't require your staff to become data scientists or spend hours generating reports. AI surfaces the right information at the right time, empowering your team to make better decisions faster—and ultimately, to focus on the strategic work that advances your organization's mission.

How FMX Is Incorporating AI: The Iris Approach

Solve Real Problems

FMX's approach to AI reflects a core principle: solve real problems first, apply technology second. This means starting with the pain points operations teams actually experience, then applying AI where it can genuinely help—not adding AI features simply because they're trendy.

Introducing Iris

Iris is FMX's AI-powered assistant, designed to serve as that "comfortable interface" between users and the robust FMX platform. Iris embodies several key principles:

Conversational by design

Rather than requiring users to master complex interfaces, Iris allows natural language interaction. Ask questions, submit requests, or search for information using the words that come naturally.

Contextually aware

Iris understands your role, your facility, and your history within the system. This context allows for more relevant suggestions and more accurate interpretations of requests.

Progressively helpful

Iris is designed to assist with simple tasks immediately while continuously learning to handle more complex workflows over time.

Security and Compliance

Public-sector organizations rightfully have stringent requirements around data security and privacy. FMX's AI implementation:

• Maintains data within secure, compliant infrastructure
• Doesn't use customer data to train public models
• Provides audit trails for AI-assisted actions
• Allows administrators to configure AI capabilities based on organizational policies
• Only apply AI where it matters

Best Practices for AI Adoption in Operations

Based on research and early adopter experiences, consider these guidelines when incorporating AI into your operations:

1. Start With High-Impact, Low-Risk Use Cases

Begin with applications where AI can deliver clear value without requiring perfect accuracy:

• Information retrieval (helping staff find documentation or asset history)
• Data entry assistance (auto-filling forms, suggesting categories)
• Report generation (creating summaries of work completed)

Remember that high-stakes decisions (like budget allocation or personnel decisions) always require human judgment. AI can support but not replace human intelligence and intuition.

2. Focus on User Adoption, Not Just Technical Implementation

The best AI capabilities are useless if your team won't use them:

• Involve frontline staff in identifying pain points and testing solutions
• Provide clear examples of how AI makes their jobs easier
• Celebrate early wins and share success stories
• Make AI features optional initially, allowing organic adoption

3. Maintain Human Oversight

AI should augment human decision-making, not replace it:

• Design workflows where AI suggests and humans approve
• Create clear escalation paths for uncertain situations
• Regularly review AI-assisted decisions to ensure quality
• Empower staff to override AI recommendations when their expertise dictates

4. Prioritize Data Quality Over Data Quantity

AI is only as good as the data it learns from:

• Start with clean, well-organized data in core areas
• Use AI to help improve data quality going forward
• Accept that you don't need perfect historical data to benefit from AI
• Focus on consistent data collection moving forward

5. Plan for Change Management

Introducing AI changes workflows and roles:

• Communicate clearly about what AI will and won't do
• Address concerns about job security directly and honestly
• Provide training that focuses on outcomes, not technology
• Frame AI as a tool that elevates roles rather than eliminates them

6. Measure Impact, Not Just Activity

Track metrics that matter:

• Time saved on administrative tasks
• Improvement in response times or completion rates
• Increase in data accuracy or completeness
• Staff satisfaction and adoption rates
• Reduction in costly emergencies due to better preventive maintenance

7. Build Internal AI Literacy

Help your team understand AI without requiring technical expertise:

• Explain what AI is good at (pattern recognition, information retrieval) and what it's not (making values-based decisions, replacing human expertise)
• Share concrete examples from your organization
• Address misconceptions and concerns openly
• Create champions who can support peers

Security, Compliance, and Ethical Considerations

Key Questions for Public-Sector AI Adoption

Before implementing AI capabilities, ensure you can answer these questions:

Data Privacy

• Where is our data stored and processed?
• Who has access to data used for AI training or inference?
• How is personally identifiable information protected?
• What happens to our data if we discontinue the service?

Compliance

• Does the AI implementation meet relevant regulatory requirements (FERPA for schools, HIPAA for healthcare facilities, etc.)?
• Are there audit trails for AI-assisted decisions?
• Can we explain how AI-driven recommendations are generated?

Fairness and Bias

• How do we ensure AI doesn't perpetuate existing biases in our operations?
• Are AI recommendations consistent across different user groups and facility types?
• What processes exist to identify and correct bias?

Transparency

• Can users tell when they're interacting with AI versus traditional software?
• Do we have clear policies about appropriate AI use?
• How do we handle situations where AI makes mistakes?

Common Concerns and How to Address Them

"Will AI replace our staff?"

The reality: AI replaces mouse clicks, not technicians and administrative staff. The goal is to automate repetitive administrative tasks so your team can focus on work that requires human judgment, relationship-building, and creative problem-solving.

The opportunity: As administrative burden decreases, your team has more capacity for strategic initiatives, stakeholder relationships, and the complex problem-solving that justifies their expertise.

"Our team isn't technical enough to use AI."

The reality: Modern AI is designed to reduce technical requirements, not increase them. If your team can have a conversation, they can use conversational AI.

The opportunity: AI can actually democratize access to technical capabilities, allowing non-technical staff to accomplish tasks that previously required specialized skills.

"We don't have enough good data to benefit from AI."

The reality: While AI needs data to generate insights, it can also help you collect better data going forward. You don't need perfect historical data to start.

The opportunity: Use AI-assisted data entry to improve data quality from this point forward. Even a few months of better data can yield meaningful insights.

"AI is too expensive or complex to implement."

The reality: AI implementation varies widely in cost and complexity. Many modern solutions, including FMX's Iris, integrate into existing workflows with minimal disruption.

The opportunity: Start small with low-cost, high-impact applications. Prove value in one area before expanding to others.

"How do we know we can trust AI recommendations?"

The reality: AI should be treated like any decision-support tool—helpful but requiring human judgment.

The opportunity: Design workflows where AI suggests and humans approve. Over time, as trust builds, you can expand the scope of AI autonomy in low-risk areas.

Call to Action: It's Foolish Not to Figure This Out

AI is not a passing trend. It's a fundamental shift in how we interact with software and leverage data. For operations leaders, the question isn't whether to engage with AI, but how to do so thoughtfully and effectively.

The organizations that will thrive are those that approach AI with:

• Clarity of purpose – focusing on real problems, not trendy technology
• Commitment to users – making systems more accessible, not more complex
• Patience with process – allowing for learning and iteration
• Confidence in value – trusting that better tools lead to better outcomes

Where to start:

1. Identify one painful, repetitive task your team handles regularly
2. Explore if AI can help by testing available features in your current tools
3. Measure the impact on time, accuracy, and staff satisfaction
4. Share the results to build momentum for broader adoption
5. Iterate and expand based on what you learn

Conclusion: The Human-Centric Future

AI Doesn't Replace Operations Staff—It Gets Them Out From Behind the Desk

The future of AI in operations isn't about autonomous buildings that run themselves. It's about empowering the people who run those buildings to work more effectively, make better decisions, and spend their time on work that matters.

When AI serves as a "comfortable interface," several things happen:

• Data gets collected because it's easy, not burdensome
• Insights emerge without requiring manual analysis
• Decisions get made with better information and less administrative overhead
• Staff feel empowered rather than overwhelmed
• Organizations demonstrate value through measurable outcomes

The Takeaway

The goal of AI in operations isn't just "smarter buildings"—it's simpler days for the people running them, coupled with better insights for the organizations depending on them.

You don't need to be a technologist to lead this transformation. You need to be what you already are: someone who understands the real challenges of operations, who knows your team and facilities, and who can identify where reducing friction will create the most value.

The AI revolution in operations isn't about the technology—it's about using that technology to make your job, and your team's jobs, focused on the work that requires your unique human expertise.

That's a future worth building.

Additional Resources

Learn More About FMX and AI

• Explore Iris capabilities within your FMX instance
• Connect with FMX customer success team to discuss AI adoption strategies
• Attend FMX webinars and user conferences for ongoing education

Questions for Further Discussion

• What repetitive tasks consume the most time for your operations team?
• Where does data collection fail in your current workflows?
• What decisions would be easier with better information at your fingertips?
• How could your team spend their time if administrative burden decreased by 20%?